1. 0:00Welcome to the DVA Atlas. Dynamic visual acuity — DVA — is the test that asks a deceptively simple question. When you move your head, does the world stay still on your retina?
2. 0:18If the vestibulo-ocular reflex is doing its job, the answer is yes. Your eyes counter-rotate at exactly the speed of your head, and the scene stays sharp. If the reflex is failing, the image slips across the retina with every step, every nod, every glance over the shoulder. The patient experiences this as oscillopsia — the world bouncing — and as deteriorating visual acuity during movement.
3. 0:50DVA quantifies that deterioration. You measure visual acuity twice: once with the head still, once with the head moving. The drop in logMAR — or in lines on a Snellen chart — is the DVA loss. Two lines or 0.2 logMAR is the most widely cited cut-off; above that, the test is considered abnormal.
4. 1:20Why does this matter? Because DVA is the only routine vestibular test that measures the functional consequence of VOR failure, not the failure itself. A caloric test tells you the canal is weak. A head-impulse test tells you the reflex gain is low. DVA tells you the patient can no longer read the highway sign while walking. That is what they came to the clinic for.
5. 1:50The test exists in three forms. Bedside DVA — a Snellen chart and the examiner's hands, oscillating the head at about two hertz. Computerised DVA — an optotype displayed only when a head-mounted rate sensor confirms the head is moving fast enough. And head-thrust DVA — single unpredictable head impulses in canal planes, probing individual semicircular canals.
6. 2:20Across this atlas we will work through anatomy, technique, normal findings, and the signatures of disease. Two reading levels above this one — Trainee and Clinician — add depth as you go. The atlas is for educational purposes only. Clinicians remain completely responsible for every clinical decision.

What is Dynamic Visual Acuity?

When you turn your head, your eyes counter-rotate to keep the image of the world steady on your retina. The reflex that does this is the vestibulo-ocular reflex (VOR). It is fast — much faster than tracking with your eyes — and it works almost entirely below conscious awareness.

Dynamic visual acuity (DVA) measures how well that reflex is doing its job. The principle is simple. First, measure how small a letter the patient can read with the head still: this is static visual acuity (SVA). Then measure the same thing while the head is moving — typically a side-to-side oscillation at about two cycles per second. The difference between the two scores is the DVA loss.¹

Why this test matters

A patient with bilateral vestibular loss may describe their problem as "the world bounces when I walk," or "I can't read the signs from the car." This is oscillopsia — the visual experience of a failing VOR. Other vestibular tests, such as the caloric or the head-impulse test, tell you that the reflex is impaired. DVA tells you how much it interferes with the patient's everyday vision.⁶

DVA is used at three points in a clinical pathway. As a screening test — picking out patients whose vestibular loss is functionally significant. As a monitoring test — tracking recovery during vestibular rehabilitation. And as a localising test — particularly in its head-thrust form, where it can implicate individual semicircular canals.^3,4

Three forms of the test

DVA exists in several closely related forms. They share the same underlying logic — read an optotype during head motion — but differ in how the head is moved, how the optotype is shown, and what claim the test can support.

1. Bedside DVA

The examiner stands behind the seated patient and oscillates the head horizontally (or vertically) at about two hertz while the patient reads a wall-mounted Snellen chart. A drop of more than two chart lines from the static baseline is taken as abnormal.^6,11 The test takes under a minute and needs no equipment beyond a chart.

2. Computerised DVA

A head-mounted rate sensor monitors angular head velocity. An optotype — typically a randomly oriented letter E — is flashed on a screen only during a brief window when head velocity is above a pre-set threshold (commonly 120°/s). The size of the smallest optotype the patient can identify is recorded; the test is repeated and an adaptive algorithm converges on threshold acuity.^1,7

The Herdman 1998 validation reported a sensitivity of 94.5% and specificity of 95.2% for identifying vestibular hypofunction against a mixed control and patient cohort.¹

3. Head-thrust DVA

A short, rapid, unpredictable head impulse in the plane of a single semicircular canal replaces the sinusoidal oscillation. Acuity is measured during the impulse window. Because the canals can be stimulated one at a time, the test can localise loss to an individual canal — useful in vestibular neuritis (where the superior or inferior division may be selectively involved) and in superior canal dehiscence surgery follow-up.⁴

Schubert and colleagues proposed a head-thrust DVA cut-off of 0.158 logMAR (mean + 2 SD in their healthy control cohort) for abnormality.⁴

Reading the result

DVA loss is most commonly expressed in logMAR units — the logarithm of the minimum angle of resolution. A logMAR of 0.0 corresponds to Snellen 20/20; each increment of 0.1 logMAR is one line on a standard chart. A DVA loss of 0.2 logMAR (two lines) is the most widely cited threshold for abnormality on the bedside test.⁶ On computerised testing, threshold criteria depend on the specific paradigm and lab norms.^1,4

Where DVA fits in the test battery

DVA is not a replacement for the caloric test or for video head-impulse testing (vHIT). It is a complementary functional measure: it asks not whether the canal is responsive, but whether the patient's visual world stays stable while they live in it. A vHIT may show a reduced gain with corrective saccades; DVA tells you whether those saccades are quick enough to recover the image before perception suffers.^3,5

The mechanism of DVA recovery during vestibular rehabilitation has been partly characterised. Schubert and colleagues (2008) attributed improvement to two changes — a small rise in active VOR gain, and an increase in the number of well-timed compensatory saccades during the head motion. The latter may dominate in chronic peripheral loss.⁵

Evidence summary

The reliability of computerised DVA has been established for both healthy controls and for patients with vestibular hypofunction across multiple studies.^1,2,3 Test-retest reliability of the active yaw protocol has been reported with ICC values around 0.83. The clinical bedside DVA has weaker psychometric properties than the computerised version but is widely used because of its accessibility and the speed with which it can be incorporated into the routine neurotologic examination.⁶

Paediatric DVA testing has been adapted by Rine and Braswell using horizontal head rotation in the yaw plane at 2 Hz; the test is reliable in children as young as three years.⁹ A vertical DVA variant exists but is generally poorly tolerated and rarely used clinically.¹⁰

What this atlas covers

The remaining modules walk through:

• Anatomy and physiology of the VOR, gaze-stabilisation circuits, and the saccadic systems that supplement the reflex when it fails.
• Technique — bedside, computerised, gaze-stabilisation test (GST), and head-thrust DVA in all six canal planes.
• Normal findings, with age-banded norms and a discussion of test-retest variability.
• Disease signatures — bilateral vestibulopathy, unilateral neuritis, Ménière's, schwannoma, vestibular migraine, central causes, presbyvestibulopathy, and ototoxicity.
• Tools and assessment — clinical cases, pattern recognition, comparison, and self-assessment.

1. 0:00The vestibulo-ocular reflex is the fastest reflex in the human body. From hair cell to extraocular muscle, the signal travels through just three neurons. Latency at the muscle is between seven and fifteen milliseconds — short enough that the eyes are already counter-rotating before the head has finished its turn.
2. 0:25The peripheral apparatus is the membranous labyrinth: three semicircular canals that sense angular acceleration, and two otolith organs — the utricle and saccule — that sense linear acceleration and tilt. For dynamic visual acuity, the canals matter most, because DVA tests the rotational VOR.
3. 0:55The canals are arranged in three orthogonal planes. The horizontal canals of the two ears form a coplanar pair. The right anterior canal is coplanar with the left posterior — that is the RALP plane. The left anterior pairs with the right posterior — the LARP plane. Head motion in any direction is decomposed into these three pairs.
4. 1:25Each canal works in push-pull with its contralateral pair. A head turn to the right increases firing in the right horizontal canal and decreases it in the left. The difference signal — not either signal alone — is what the vestibular nuclei integrate.
5. 1:55From there, the three-neuron arc is simple. Primary afferent in the Scarpa's ganglion. Second-order neuron in the vestibular nucleus — predominantly superior and medial for canal signals. Third-order motor neuron in the ocular motor nuclei — third, fourth, or sixth nerve.
6. 2:25When the reflex fails, the eyes lag the head. The image slips across the retina; the patient sees the world bouncing. To recover, the brain produces catch-up saccades — covert if they happen during the head motion, overt if after. Covert saccades preserve dynamic visual acuity. Overt saccades cannot; they arrive too late.
7. 2:55This is what DVA measures. Not whether the reflex exists, but whether the eyes are fast enough — by reflex or by saccade — to keep the world readable while the head is moving.

Why the VOR exists

A clear retinal image is impossible if the eye moves with the head. Walk a few steps holding a book and try to read it without moving your eyes at all — the print becomes a blur. The vestibulo-ocular reflex solves this problem by rotating the eyes in the opposite direction to the head, at the same speed, with a latency of only seven to fifteen milliseconds.¹⁶

That latency is what makes the VOR special. Smooth pursuit eye movements need around 100 ms to start and can only track targets moving slowly. The VOR runs through three synapses in the brainstem, fast enough that the eyes are already moving by the time conscious awareness of the head turn registers.^13,16

The peripheral apparatus

Each inner ear contains a membranous labyrinth — five fluid-filled sensory organs. For DVA we are interested chiefly in the three semicircular canals (horizontal, anterior, and posterior) which detect angular acceleration. The two otolith organs (utricle and saccule) detect linear acceleration and tilt; they contribute to the translational VOR but play a smaller role in the test as routinely performed.¹²

The canals are arranged so that each ear's canal is coplanar with one canal on the other side: horizontal pairs with horizontal, the right anterior pairs with the left posterior (the RALP plane), and the left anterior pairs with the right posterior (the LARP plane). The brain reads the difference in firing rate between each coplanar pair, not the firing of either canal alone.^12,13

The three-neuron arc

The classical VOR arc has exactly three neurons between sensor and effector — that is the source of its speed.¹³

1. Primary afferent. The hair cells in the cristae of the semicircular canals synapse on bipolar neurons whose cell bodies lie in Scarpa's ganglion. Their axons form the vestibular division of cranial nerve VIII.
2. Vestibular nucleus neuron. The afferent terminates in the vestibular nuclear complex in the lateral medulla. Canal inputs synapse predominantly in the superior and medial vestibular nuclei.¹²
3. Ocular motor neuron. The second-order neuron projects — most often across the midline, sometimes ipsilaterally — to the abducens (VI), oculomotor (III), or trochlear (IV) nucleus, which drives the appropriate extraocular muscle.

For a horizontal head turn to the right (as drawn above), the right horizontal canal's firing rate increases. The right vestibular nucleus excites the left abducens nucleus, which drives the left lateral rectus directly through VI and the right medial rectus indirectly through the medial longitudinal fasciculus and the oculomotor nucleus. Both eyes rotate leftward — opposite the head turn — and gaze stays fixed.¹³

Canal-to-muscle wiring

Each semicircular canal drives a specific pair of extraocular muscles. For DVA testing — especially head-thrust DVA — knowing which canal pair corresponds to which muscle pair lets you interpret an asymmetric result topographically. The pairings are the same as for video head-impulse testing.¹⁴

The push-pull principle

At rest, vestibular afferents fire spontaneously at roughly 90 spikes per second. A head turn modulates this baseline — the canal on the side of the turn fires faster, the contralateral canal fires slower. The vestibular nuclei compare the two firing rates. Symmetric spontaneous firing makes excitation possible in either direction from baseline.^12,13

This is why an acute unilateral vestibular lesion produces such a dramatic clinical picture. The intact side keeps firing at 90 spikes per second; the lesioned side falls to zero. The brain reads the difference — a tonic asymmetry equivalent to a permanent head turn toward the intact ear — and produces spontaneous nystagmus until central compensation rebalances the system.

Gain, phase, and what they mean

The gain of the VOR is the ratio of eye velocity to head velocity. A perfect compensatory reflex has a gain of 1.0 (or −1.0 if you adopt the sign convention that eye and head move in opposite directions). Healthy adults achieve gain of approximately 0.95–1.0 for high-frequency, high-velocity stimuli such as head impulses.¹⁴

Phase describes the timing relationship — whether eye velocity precedes or lags head velocity. At physiological frequencies (2–20 Hz) phase is essentially zero; eye and head move together. Phase abnormalities matter mostly for low-frequency rotational chair testing and have limited bearing on routine DVA interpretation.

Catch-up saccades

When VOR gain falls, the eyes lag the head; gaze position error accumulates during the head motion. The brain recovers gaze with catch-up saccades — rapid corrective eye movements directed at the visual target.¹⁴

• Covert saccades occur during the head motion, typically with latencies under 200 ms. Because they happen while the head is still moving, they can recover the image before it slips far enough off the fovea to blur acuity. Covert saccades are the principal mechanism by which DVA recovers after vestibular rehabilitation.⁵
• Overt saccades occur after head motion has ended. They are slower to programme and arrive too late to preserve dynamic visual acuity — the patient has already missed the optotype by the time the eye catches up to the target.¹⁴

Two patients with identical reduced VOR gain can have very different DVA scores depending on whether they generate covert or overt catch-up saccades. Older adults compensate with larger but appropriately-timed saccades that partly preserve dynamic vision despite age-related gain loss.¹⁵

Frequency response of the VOR

The VOR is most sensitive at the high-frequency, high-velocity end of the spectrum — the regime in which a single brisk head turn occurs. Caloric testing, by contrast, probes the canal at frequencies around 0.003 Hz — three orders of magnitude below physiological head motion. The two tests therefore interrogate different operating regions of the same organ, and can dissociate: an absent caloric with a present head-impulse response is well documented, and the converse exists too.^14,16

DVA at 2 Hz oscillation falls in the same high-frequency regime as the head-impulse test, which is why the two tests tend to agree in direction and why DVA can be interpreted as a functional consequence of the vHIT abnormality.^6,16

The otolith contribution

The translational VOR — driven by the utricle and saccule — contributes during walking, where each step delivers a brief vertical translation of the head. The translational reflex stabilises gaze at near distances; the rotational reflex dominates at far. For routine DVA, the target is far (typically two metres) and the rotational contribution dominates, so the test is comparatively insensitive to isolated otolith dysfunction.¹³

1. 0:00DVA testing comes in four closely related forms. All four share the same logic — read an optotype during head motion — but they differ in how the head is moved, how the optotype is shown, and what claim the test can support.
2. 0:20The bedside test is the simplest. The patient sits in front of a Snellen chart at two to three metres. Static visual acuity is established first — the smallest line the patient can read with the head still. Then the examiner stands behind the patient, grips the head firmly at the temples, and oscillates it horizontally at about two cycles per second. The patient reads down the chart. A drop of more than two lines from static is abnormal.
3. 0:55Two practical points for the bedside test. First, the oscillation must be passive — the examiner moves the head, not the patient. Active motion lets the brain preprogramme compensatory eye movements and inflates the score. Second, the frequency must be at least two hertz. Below that, smooth pursuit can keep up with the chart and the VOR is not specifically tested.
4. 1:25Computerised DVA replaces the human examiner's eye with a rate sensor and an optotype-flashing computer. The patient wears a head-mounted gyroscope. An optotype — typically a randomly rotated letter E — appears on the screen only when head velocity exceeds a threshold. The Herdman protocol uses 120 degrees per second; the NIH Toolbox protocol uses 180. Optotype dwell is around 80 milliseconds — short enough to exclude pursuit.
5. 2:00The gaze stabilisation test inverts the question. Instead of asking how small an optotype the patient can read at a fixed velocity, it asks how fast the head can move while the patient still reads a fixed optotype. The outcome measure is peak head velocity, not visual acuity. Reduced gaze stabilisation velocity on one side compared to the other is the signature of unilateral vestibular loss.
6. 2:35Head-thrust DVA is the most demanding paradigm. A single passive, rapid, unpredictable head impulse replaces the sinusoidal oscillation. The impulse is delivered in the plane of a single semicircular canal — horizontal for the lateral canals, RALP for the right anterior and left posterior, LARP for the left anterior and right posterior. Because the canals are stimulated one at a time, the test localises loss to individual canals.
7. 3:10Use the interactive simulator on this page to feel how gain, frequency, and amplitude interact. The simulator is a teaching caricature, but the qualitative relationship — slip rises as gain falls, and acuity drops with slip — is the substrate of every clinical DVA test you will perform.

The DVA simulator

Three sliders — VOR gain, head frequency, head amplitude — drive a live optotype and a head silhouette. Watch what happens to retinal slip as gain drops, and how the optotype blurs as slip rises. The numeric read-out estimates the steady-state DVA loss in logMAR.

1. Bedside DVA

The bedside test requires only a wall-mounted Snellen chart. Static visual acuity is established first — the smallest line the patient can read with the head still. The examiner then stands behind the seated patient, grips the head firmly at the temples, and oscillates it horizontally at approximately two cycles per second. The patient reads down the chart. A loss of more than two lines from the static baseline is taken as abnormal.^6,11

Step-by-step

1. Seat the patient at two to three metres from a wall-mounted Snellen chart, with corrective lenses worn for distance if normally used.
2. Record static visual acuity — the smallest line the patient can read with the head still. This is the reference.
3. Stand behind the patient, grip the head firmly at the temples or occiput, and oscillate horizontally at approximately 2 Hz with peak displacement of ±20–30 degrees. Avoid pausing at the turn-arounds.
4. Ask the patient to read down the chart. Record the smallest line they can identify during oscillation.
5. Calculate the lines lost. ≥3 lines is abnormal; 2 lines is borderline and warrants further testing.⁶

2. Computerised DVA

The computerised test replaces the human examiner with a head-mounted rate sensor and a screen that flashes optotypes only when head velocity exceeds a pre-set threshold. The Herdman protocol uses ≥120°/s; the NIH Toolbox uses ≥180°/s with an 83 ms optotype dwell — short enough to exclude smooth pursuit and pre-programmed saccades.^1,18

A randomly oriented letter E (or Landolt C) appears briefly when the velocity gate triggers. The patient identifies its orientation; an adaptive algorithm converges on the smallest reliably identified size. Static visual acuity is measured first, dynamic acuity is measured per direction (rightward and leftward head motion separately), and the per-direction loss in logMAR is reported.⁷

Normative values

The NIH Toolbox normative study (n = 3,992, ages 3–85) is the largest available reference dataset. cDVA was worse in males than females, and began to decline from age 50 onwards; younger age bands (3–49) showed no significant difference.¹⁸ Most clinical labs use a per-direction loss of ≥0.2 logMAR as the cut-off for abnormality, with asymmetry of ≥0.1 logMAR between sides considered significant.

3. Gaze stabilisation test (GST)

The GST inverts the DVA question. Instead of asking how small an optotype the patient can read at a fixed velocity, it asks how fast the head can move while the patient still reads a fixed optotype. The outcome is peak head velocity at which the patient reliably identifies the optotype, reported per direction.¹⁷

A fixed optotype size — typically 0.2 logMAR above the patient's static acuity — is used.²⁰ The patient performs head motion at progressively increasing velocity bands (60–99, 100–139, 140–179, 180–219, ≥220°/s in the original Goebel protocol). The test stops when the patient can no longer identify the optotype at three out of five presentations in the current band; the maximum velocity for the preceding successful band is recorded.¹⁷

GST values vary substantially across labs because protocol details differ — testing distance, optotype size, the increment between velocity bands. Reduced contralesional GST velocity is a sensitive marker of unilateral vestibular dysfunction; an asymmetry index of ≥25% between sides is commonly considered abnormal.^17,20

4. Head-thrust DVA (htDVA)

The most demanding paradigm. A single passive, rapid, unpredictable head impulse — in the plane of a specific semicircular canal — replaces sinusoidal oscillation. Acuity is measured during the impulse window. Because the canals can be stimulated one at a time, the test localises loss to individual canal pairs.⁴

Schubert and colleagues proposed a head-thrust DVA cut-off of 0.158 logMAR (mean + 2 SD in their healthy control cohort) for abnormality.⁴ The impulse must be brisk: an acceleration of 3,000–6,000°/s² is typical, with peak velocity ≥200°/s.

Which test, when?

The four paradigms are not interchangeable. They probe the VOR at different operating points and report different things, so the right test depends on the clinical question.^7,19

• Bedside DVA — for the initial screen in any patient with dizziness, oscillopsia, or imbalance. Fast, free, and 80% sensitive for bilateral vestibular loss in experienced hands. Not sensitive enough for unilateral subtle loss.⁶
• Computerised DVA — for quantitative documentation (research, medico-legal, rehabilitation monitoring). Reproducible (ICC ~0.8) and well-normed across ages.^18,19
• GST — for patients whose static DVA is normal but who complain of oscillopsia at high head velocities. The GST extends the test into a higher-velocity regime than standard cDVA.¹⁷
• htDVA — for topographic questions. Selective inferior-division neuritis, post-SCD plugging follow-up, partial canal palsy. Pair with video head-impulse testing.⁴

Technical considerations

Three engineering details that disproportionately affect test performance, in order of impact:

1. Velocity gating. If the optotype is displayed at head velocities below ~120°/s, smooth pursuit and pre-programmed saccades contribute and the test loses VOR specificity. Set the threshold high.^2,18
2. Optotype dwell. Display times under 100 ms exclude pre-programmed compensatory saccades, which have a minimum latency of about 100 ms. The NIH Toolbox standard of 83 ms is a sound default.¹⁸
3. Optotype family. Letter E and Landolt C tests require the same orientation judgement and are functionally equivalent. Mixed letter sets (ETDRS) increase difficulty but allow normative comparison with ophthalmic acuity testing.¹⁸

Using DVA to monitor vestibular rehabilitation

DVA improves with vestibular rehabilitation in both unilateral and bilateral loss.^3,5 Schubert and colleagues (2008) attributed the improvement to two mechanisms: a small rise in active VOR gain and an increase in the number of well-timed compensatory saccades during the head motion. The latter often dominates in chronic peripheral loss.⁵

When monitoring a rehabilitation programme, the same paradigm should be used at each visit — bedside is not interchangeable with computerised. A loss reduction of 0.1 logMAR (one chart line) over four to six weeks of structured gaze-stabilisation exercise is a typical responder pattern.⁵

1. 0:00Before you can recognise an abnormal DVA, you have to know what normal looks like across age and across paradigms. The largest reference dataset comes from Li, Beaumont, Rine, Slotkin, and Schubert in 2014 — the NIH Toolbox normative study of nearly four thousand individuals aged three to eighty-five.
2. 0:25Two findings dominate. First, between ages three and forty-nine, there is no significant change in DVA performance. A nine-year-old and a thirty-five-year-old score similarly. Second, from age fifty onwards, performance declines progressively. By the seventies and eighties, the upper limit of normal has widened well beyond the 0.2 logMAR threshold used at the bedside.
3. 0:55This matters because applying a single bedside threshold to a seventy-five-year-old can falsely label them abnormal. The age-matched mean plus two standard deviations is the right reference for the computerised test, and most clinical labs maintain their own age-banded norms.
4. 1:25Sex matters slightly. Males score worse than females in the NIH Toolbox dataset, a small but statistically significant difference. Ethnicity and education show no meaningful effects after adjusting for the visual acuity floor.
5. 1:50Test-retest reliability of computerised DVA is high. Intraclass correlations cluster around 0.83 for the active yaw protocol. The bedside test is less reliable — examiner-dependent variation in head motion frequency and amplitude introduces noise. Reproducibility over weeks is good in healthy subjects and poor in active disease, which is itself diagnostically useful.
6. 2:25Three practical reminders. Always pair a static visual acuity measurement with the dynamic one — DVA loss is the difference, not the absolute. Always test both directions of head motion. And always interpret the result against the right paradigm-specific and age-banded cut-off.

What normal looks like

In a healthy adult under fifty, DVA loss is small — typically under 0.1 logMAR (one line on a Snellen chart) at the standard bedside oscillation frequency of two hertz. The NIH Toolbox normative study, the largest available dataset, reported a population mean DVA loss across all ages of 0.116 ± 0.184 logMAR (n = 3,992).²¹

That mean conceals an important age effect. Between ages three and forty-nine, performance is essentially flat; a child and a middle-aged adult score similarly. From age fifty onwards, DVA loss climbs progressively, and the upper limit of the normal range widens accordingly.^21,23

Static visual acuity comes first

DVA loss is a difference, not an absolute. Static visual acuity must be measured first; the dynamic measurement is interpreted relative to it. A patient with uncorrected refractive error or cataract may have a low static acuity that produces a floor effect — the smallest optotype they can read with the head still is already large enough that head motion makes little additional difference, and the DVA loss looks falsely normal.

Correct vision first. Reading glasses worn for distance during the test. Then measure static, then dynamic. The Herdman 1998 protocol, the NIH Toolbox protocol, and the Schubert head-thrust paradigms all enforce this order for the same reason.^1,7,18

Age effects in detail

The Li 2014 dataset combined the paediatric population (ages 3–17) into a single band and analysed adults in decade-wide bands thereafter. Three findings deserve emphasis.²¹

1. No paediatric handicap. Three-year-olds with normal vestibular function score similarly to teenagers and young adults. This makes DVA usable as a screening test in children — a rare attribute among vestibular tests, most of which require adult cooperation. The Rine paediatric protocol carries 100% sensitivity and high specificity in children with bilateral vestibulopathy.⁹
2. A plateau through middle age. Between 18 and 49, the mean and SD are essentially stable. A 35-year-old with a DVA loss of 0.2 logMAR is borderline regardless of which decade of adulthood they sit in.
3. Decline from fifty onwards. By the 60–69 band, the mean has roughly doubled; by 70–85 it is approaching the bedside threshold even in healthy individuals. The SD widens in parallel, so the +2 SD upper limit of normal moves up faster than the mean itself.²¹

Sex and other factors

Males scored worse than females in the NIH Toolbox dataset (p < 0.001). The magnitude is small but statistically significant; the effect was present at every age band.²¹ Ethnicity, dominant language, and education showed no meaningful effects after adjustment.

Refractive error and visual function should be corrected before testing. Cervical spine instability, severely limited neck range of motion, and oculomotor impairment that prevents stable fixation are exclusion criteria — the test cannot validly interrogate the VOR if the head or eyes cannot do what the protocol requires.²¹

Reliability and reproducibility

Computerised DVA is reproducible to a degree that makes it suitable for monitoring response to vestibular rehabilitation. The Herdman 1998 protocol reported test-retest ICCs around 0.83 for the active yaw paradigm.^1,19 Mohammad et al. (2011) confirmed that both cDVA and GST are stable across visits in patients with peripheral vestibular disorders.¹⁹

Bedside DVA is less reproducible. Examiner-dependent variation in head oscillation frequency and amplitude introduces noise that the computerised paradigm eliminates. For monitoring change over time — rehabilitation, post-surgical recovery, ototoxic risk surveillance — stick with the computerised test and the same operator if possible.⁶

Choosing the right cut-off

Three reasonable strategies exist for declaring a cDVA score abnormal, and a clinician should know which their lab uses.^1,21

• Fixed cut-off — 0.2 logMAR per direction. The Herdman 1998 paradigm cut-off, equivalent to two chart lines. Simple, widely-used, but over-calls older adults.¹
• Age-matched mean + 2 SD. Preferred when an age-banded normative dataset is available. The Li 2014 dataset is the largest published reference for the NIH Toolbox protocol; lab-specific norms should otherwise be derived from at least 30 controls per age band.²¹
• Asymmetry of ≥ 0.1 logMAR between directions. Reads the test against itself, so it is robust to age effects and to floor effects in the static measurement. The most useful criterion for screening unilateral vestibular loss; insensitive to symmetric bilateral loss.⁷

Diagnostic accuracy at the published thresholds

For computerised DVA against a reference battery of caloric testing and rotational chair, the Herdman 1998 paradigm reported sensitivity 94.5% and specificity 95.2% for identifying any vestibular hypofunction in a mixed cohort.¹ The Vital 2010 protocol using 100°/s and 150°/s active yaw impulses reported 100% sensitivity against scleral search-coil VOR gain measurement.²²

In the paediatric population, Rine and Braswell 2003 reported 100% sensitivity and specificity for identifying bilateral vestibulopathy in children with sensorineural hearing impairment.⁹These published values come from research-grade protocols and well- characterised cohorts; real-world clinic performance is typically lower, particularly for unilateral subtle loss.

DVA in the context of the full vestibular battery

A normal DVA does not rule out vestibular pathology. The test probes the high-frequency VOR; isolated otolith dysfunction, low-frequency caloric weakness, and well-compensated chronic unilateral loss can all present with normal DVA scores.¹³ The right interpretive frame is:

• DVA abnormal, vHIT abnormal → high-frequency VOR loss with functional consequence; classic peripheral vestibular hypofunction pattern.
• DVA abnormal, vHIT normal → consider central origin, oculomotor disorder, or visual processing problem. Recheck refractive correction.
• DVA normal, vHIT abnormal → well-compensated peripheral loss, often with effective covert catch-up saccades. Patient may still be symptomatic at extremes of head velocity (GST helps here).^5,17
• DVA normal, vHIT normal, caloric abnormal → low-frequency canal loss only. Hydrops should be considered.

DVA during walking and passive translation

Routine clinical DVA is performed seated, with rotational head motion. Walking-DVA — performed on a treadmill while the patient reads a chart — interrogates the translational VOR in addition to the rotational, and is more sensitive to chronic bilateral loss than seated DVA. It is rarely available outside research settings.¹⁹

A version of DVA using passive vertical motion of the patient on an oscillating chair has been described for the bilaterally hyporeflexic population. It picks up patients in whom seated rotational DVA appears falsely normal because of robust efference-copy preprogramming. Specialised laboratories only.^19,22

1. 0:00Bilateral vestibulopathy is the textbook indication for the DVA test. It is the disease in which DVA earns its keep more than any other vestibular condition — and the only one in which a normal subjective visual vertical paradoxically coexists with profound vestibular impairment.
2. 0:25The Bárány Society defines bilateral vestibulopathy as a chronic vestibular syndrome with unsteadiness when walking or standing, worse in darkness or on uneven ground, and head-motion-induced blurred vision or oscillopsia. Crucially, there are no symptoms while sitting or lying still. The complaint is dynamic.
3. 0:55Quantitatively, the diagnosis requires bilaterally impaired or absent VOR. The thresholds: horizontal vHIT gain below 0.6 on both sides; or the sum of caloric peak slow-phase velocities less than six degrees per second per side; or rotational chair gain below 0.1 at 0.1 hertz. Any of these is sufficient. A bilaterally pathological bedside head impulse test alone defines probable bilateral vestibulopathy.
4. 1:30DVA is listed as a complementary test. The Bárány criteria specify a DVA loss of at least 0.2 logMAR — two chart lines — as pathological. In practice, severe bilateral vestibulopathy produces DVA losses of half a logMAR or more.
5. 2:00Why is the asymmetry pattern unusual? Because there is no asymmetry. Both sides are equally affected. The right-versus-left directional difference — the criterion most labs use to flag a unilateral DVA finding — is zero. Both directions are equally bad. This is one of the few clinical scenarios in which a normal directional asymmetry index is itself the diagnostic clue.
6. 2:35Aetiology is most commonly aminoglycoside ototoxicity — gentamicin is the worst offender — followed by sequelae of meningitis, bilateral Ménière's disease, and the CANVAS syndrome of cerebellar ataxia, neuropathy, and vestibular areflexia. About half remain idiopathic.
7. 3:00Recovery of vestibular function is uncommon. But DVA recovers with vestibular rehabilitation, primarily through the brain learning to deploy centrally-programmed compensatory saccades during head motion — covert saccades — even when VOR gain does not improve. DVA is therefore the right outcome measure for monitoring rehabilitation in this population.

What is bilateral vestibulopathy?

Bilateral vestibulopathy (BVP) is a chronic syndrome in which both inner ears have lost their vestibular function — or so much of it that the vestibulo-ocular reflex can no longer keep gaze stable during head motion. Patients are typically symptom-free at rest, and their complaints emerge only with movement.²⁴

The two cardinal symptoms are unsteadiness walking or standing — worse in darkness or on uneven ground, where the patient cannot use vision or proprioception to compensate — and oscillopsia, the experience of the world moving on the retina during head motion. Patients describe it as the print jumping while they read, signs blurring as they walk, or the inability to recognise faces from a moving car.^24,27

How common is it?

Bilateral vestibulopathy is uncommon. The Ward 2013 analysis of the US National Health Interview Survey estimated a population prevalence of approximately 28 per 100,000 adults.²⁷ In tertiary dizziness clinics it accounts for roughly 1% of new referrals. The condition is associated with falls, reduced functional independence, and a high disease burden.²⁷

Bárány Society diagnostic criteria

The 2017 Bárány Society consensus defines BVP as the combination of clinical features above plus laboratory evidence of bilaterally impaired or absent VOR. The quantitative thresholds are reproduced below, drawn from the consensus document.²⁴

• vHIT or scleral coil:horizontal aVOR gain < 0.6 on both sides at angular velocity 150–300°/s, and/or
• Bithermal caloric test:sum of maximal slow-phase velocities (warm + cold) < 6°/s on each side, and/or
• Rotational chair:horizontal aVOR gain < 0.1 at 0.1 Hz sinusoidal stimulation (V_max 50°/s), and/or phase lead > 68° (time constant < 5 s).

Any one of these thresholds, in the presence of the cardinal symptoms, is sufficient for the diagnosis. Probable BVP is defined by the cardinal symptoms plus a bilaterally pathological bedside head-impulse test.²⁴

DVA appears in the criteria as a complementary test: a DVA decrease of ≥0.2 logMAR is considered pathological. cVEMP, oVEMP, and the Romberg are mentioned but not included in the definition. The consensus is careful to note that DVA documents the functional consequence of the VOR loss rather than the loss itself.²⁴

Aetiology

The single most commonly identified cause is aminoglycoside ototoxicity — gentamicin in particular, with streptomycin, tobramycin, and amikacin following. Aminoglycoside vestibulotoxicity can be profound and irreversible; it characteristically spares hearing initially while devastating both vestibular ends.²⁶

Other recognised causes include bilateral Ménière's disease, meningitis (bacterial or carcinomatous) with bilateral labyrinthine involvement, autoimmune inner ear disease, bilateral vestibular schwannomas (NF2), the CANVAS syndrome (cerebellar ataxia, neuropathy, vestibular areflexia syndrome), and head trauma. Approximately half of cases remain idiopathic; in these, a neurodegenerative process is presumed.^24,26

The DVA pattern in detail

In bilateral vestibulopathy, DVA loss is large and the directional asymmetry is small. This is the inverse of unilateral vestibulopathy, where the asymmetry between rightward and leftward head motion is often the most striking finding while the absolute loss may be modest after compensation.

Two clinical implications follow. First, an asymmetry-based screening rule will miss bilateral vestibulopathy. Labs that flag only right-vs-left differences of ≥0.1 logMAR will report this patient as normal. Absolute DVA loss against the age-matched normative mean is the right criterion in this population.^21,24

Second, the magnitude of DVA loss can be large enough that the standard chart "floor" interferes. A patient losing six lines bilaterally cannot identify any optotype on a typical chart during oscillation. In severely affected patients, the test is reported as "no optotypes identified" rather than a quantitative logMAR — but this is itself a positive finding.

Why the SVV is normal

A patient with profound bilateral vestibulopathy may present with a completely normal subjective visual vertical. This is sometimes mistaken for evidence against the diagnosis. It is not.

SVV measures the asymmetry between the two graviceptive inputs. When both utricles fail symmetrically, the asymmetry is zero and the SVV is therefore normal. The patient is still profoundly impaired; the test is simply the wrong tool for symmetric bilateral loss. DVA and the head-impulse test detect what the SVV cannot.²⁴

Rehabilitation and the DVA outcome measure

Vestibular function in established BVP rarely recovers — peripheral hair cells do not regenerate in humans, and central compensation has fewer asymmetry signals to work with than in unilateral loss. Yet DVA does improve with rehabilitation.²⁵

Herdman and colleagues (2007) randomised thirteen BVP patients to structured gaze-stabilisation exercises versus placebo. The exercise group showed significant DVA improvement; the placebo group did not. Slow-phase eye velocity gain did not change in most patients — arguing strongly that improvement is mediated not by VOR-gain recovery but by the brain learning to deploy centrally- programmed compensatory saccades during head motion. Covert catch-up saccades, when well-timed, can preserve dynamic acuity even in the absence of a functioning reflex.^5,25

The clinical implication: DVA is the right outcome measure to track rehabilitation in BVP, even though caloric and vHIT results do not change. A loss reduction of 0.1 logMAR (one chart line) over four to six weeks of structured exercise is a typical responder pattern. Use the same paradigm at each visit — bedside scores are not interchangeable with computerised scores.^5,25

Differential diagnosis

The principal differential is between BVP and other syndromes that produce oscillopsia or unsteadiness:

• Cerebellar disease — typically produces gaze-evoked nystagmus, dysmetric saccades, and impaired smooth pursuit. DVA may be abnormal centrally rather than peripherally; vHIT is often preserved.¹⁶
• Sensory ataxia from large-fibre neuropathy — unsteadiness worsens in darkness as in BVP, but the head-impulse test, DVA, and caloric responses are all normal. Romberg is markedly positive; vibration sense is impaired.
• Presbyvestibulopathy — milder version of the same picture in older adults, by definition with bilaterally reduced but not absent vestibular function. The 2019 Bárány criteria for presbyvestibulopathy use a less severe vHIT-gain threshold than BVP.
• CANVAS — the cerebellar ataxia + sensory neuronopathy + vestibular areflexia syndrome. The vestibular picture is identical to other BVP; the cerebellar and neuropathic features distinguish it. Look for chronic cough, an early CANVAS clue.²⁶

Reading the report

A bilaterally symmetric DVA loss of ≥0.4 logMAR with bilaterally abnormal vHIT and caloric responses is sufficient for the diagnosis when the cardinal symptoms are present. The clinician should then pursue aetiology with a careful drug history (aminoglycosides including past gentamicin or chemotherapy with platin agents), screening audiometry, and — when the gait or neurological examination is atypical — neuroimaging and CANVAS screening (RFC1 expansion testing where available).^24,26

1. 0:00Vestibular neuritis is the unilateral counterpart to bilateral vestibulopathy. In bilateral disease the DVA pattern is symmetric loss with no directional asymmetry. In neuritis the pattern inverts — directional asymmetry is the diagnostic clue, and the absolute loss may be modest once central compensation kicks in.
2. 0:25The Bárány Society in 2022 renamed the condition acute unilateral vestibulopathy — AUVP — keeping vestibular neuritis as an acceptable synonym. The new criteria require an acute or subacute spinning or non-spinning vertigo of moderate or severe intensity lasting at least 24 hours, spontaneous peripheral nystagmus appropriate to the canal afferents involved, unambiguous evidence of reduced VOR function on the side opposite the fast phase, and the absence of acute central or acute audiological symptoms.
3. 1:00Anatomy matters here in a way it doesn't for bilateral disease. The vestibular nerve has two divisions. The superior division carries the horizontal and anterior canals, plus the utricle. The inferior division carries the posterior canal and the saccule. Neuritis can be selective.
4. 1:25Superior division neuritis — the most common pattern — produces abnormal horizontal vHIT, abnormal caloric on the affected side, abnormal oVEMP, and preserved cVEMP. DVA shows asymmetric loss greater on head turns toward the lesioned side.
5. 1:55Inferior division neuritis is rare but instructive. Calorics and horizontal vHIT are normal — these test only the superior nerve. The posterior canal vHIT shows impulse deficit. The cVEMP, which depends on the inferior nerve, is absent. Patients can be told they have no vestibular lesion when in fact they do. Canal-specific head-thrust DVA is what catches it.
6. 2:30Recovery follows a predictable course. The acute symptoms — the spinning vertigo and the spontaneous nystagmus — subside over days to weeks as static compensation occurs in the vestibular nuclei. DVA recovers more slowly and incompletely. The Herdman 2003 trial showed that DVA improves with structured gaze-stabilisation exercise mainly through covert catch-up saccades, not through recovery of VOR gain.
7. 3:00Two practical points. First, an asymmetric DVA finding without documented unilateral VOR loss is not yet vestibular neuritis. Document the loss with caloric, vHIT, or both. Second, a normal-caloric patient with asymmetric DVA, ipsilateral hearing-spared vertigo, and absent cVEMP is inferior division neuritis. Look for it specifically.

What is vestibular neuritis?

Vestibular neuritis — also known as acute unilateral vestibulopathy (AUVP) — is an acute peripheral vestibular syndrome caused by sudden loss of vestibular function on one side, without hearing change. The patient presents with severe, sustained vertigo and unsteadiness, typically lasting days to a couple of weeks. The leading hypothesis is reactivation of latent herpes simplex virus type 1 in the vestibular ganglion, though several other mechanisms are recognised.²⁸

On examination there is a spontaneous peripheral vestibular nystagmus beating away from the affected side, made more obvious by removing visual fixation (Frenzel glasses, or M-shaped finger occlusion). Bedside head-impulse testing shows catch-up saccades when the head is thrust toward the affected ear. Hearing is preserved — this is the key distinction from labyrinthitis.²⁸

How common is it?

Vestibular neuritis is the second most common peripheral cause of acute prolonged vertigo after BPPV. The annual incidence is approximately 3.5 per 100,000 persons; the condition accounts for roughly 7% of patients presenting to vertigo clinics.²⁸

Bárány Society 2022 diagnostic criteria

The 2022 Bárány consensus introduced four diagnostic categories. The principal one is Acute Unilateral Vestibulopathy (AUVP), requiring all six criteria below.²⁸

1. Acute or subacute onset of sustained spinning or non-spinning vertigo of moderate or severe intensity, lasting at least 24 hours.
2. Spontaneous peripheral vestibular nystagmus appropriate to the canal afferents involved — generally horizontal-torsional, direction-fixed, enhanced by removal of visual fixation.
3. Unambiguous evidence of reduced VOR function on the side opposite the direction of the fast phase of the spontaneous nystagmus.
4. No evidence for acute central neurological, otological, or audiological symptoms (no hearing loss, no tinnitus, no otalgia).
5. No acute central neurological signs — no skew deviation, no gaze-evoked nystagmus, no audiological signs.
6. Not better accounted for by another disease or disorder.

The other three categories are AUVP in evolution (symptoms ≥3 hours but not yet 24 hours; used to enable specific treatment and central-stroke exclusion in the very acute window), probable AUVP (criteria met except the unilateral VOR deficit is not clearly documented), and history of AUVP (for patients seen long after the acute phase with documented unilateral VOR deficit).²⁸

Why one division gets hit more often

The vestibular nerve splits into a superior and an inferior division inside the internal auditory canal. The superior division carries afferents from the horizontal canal, the anterior canal, and the utricle; the inferior carries afferents from the posterior canal and the saccule. The two divisions take different paths through bony channels of slightly different size — the superior travels through a narrower, longer canal, which is the leading anatomical explanation for why neuritis preferentially affects the superior division.^29,31

Aw and colleagues studied 33 patients with acute unilateral peripheral vestibulopathy and found that most had abnormal lateral and anterior canal function — consistent with selective superior division involvement — while a small minority had all three canals affected, suggesting involvement of both divisions. Isolated inferior division involvement (normal calorics, abnormal posterior canal vHIT, absent cVEMP) is rare but recognised.^29,30

The DVA pattern in detail

In acute unilateral vestibulopathy, DVA is asymmetric: the loss is greater on head turns toward the affected ear than away from it. The rule comes from the canal-to-muscle wiring covered in the Anatomy module — a head turn to the right depends on the right horizontal canal driving the contralateral abducens. If the right canal is knocked out by neuritis, rightward head motion has nothing to drive the leftward eye rotation, the eyes lag, and the optotype blurs. Leftward head motion is largely intact, because it depends on the unaffected left horizontal canal.¹⁴

The clinical implications:

• Asymmetry ≥0.1 logMAR between directions is the most useful screening criterion in this population. It is robust to age, to floor effects in static acuity, and to bilateral symmetric noise. A patient with a unilateral lesion will fail this asymmetry test even when the absolute loss looks modest.^1,21
• Absolute loss on the affected side typically falls in the 0.2–0.5 logMAR range in the acute phase. Central compensation reduces the absolute loss over weeks; the asymmetry tends to persist longer than the absolute deficit and is therefore the more durable laboratory clue in the chronic phase.^3,5
• The unaffected side is rarely entirely normal in the acute phase. Patients may be too dizzy or nauseated to perform the test reliably; the bedside oscillation also affects neck and visual systems that themselves are stressed. Repeat the test once the patient can tolerate it.

Compensation and recovery

The acute symptoms — the spinning vertigo and spontaneous nystagmus — subside over days to weeks as static compensation occurs in the vestibular nuclei. The brain re-equalises the tonic firing rate between the two vestibular nuclei despite the persisting peripheral asymmetry. By two to three weeks, the spontaneous nystagmus is usually absent on direct gaze.²⁸

DVA recovers more slowly and incompletely. The Herdman 2003 trial randomised 21 patients with unilateral vestibular hypofunction to structured vestibular exercises versus placebo; the exercise group showed significant DVA improvement on the unpredictable head-motion paradigm.³ The Schubert 2008 mechanistic follow-up attributed improvement chiefly to covert catch-up saccades— well-timed, sub-200 ms saccades during the head motion — and only partly to a small rise in active VOR gain.⁵

Differential diagnosis

The most important miss is a posterior-circulation stroke presenting as an acute vestibular syndrome. The HINTS exam (Head-Impulse, Nystagmus, Test of Skew) is the bedside discriminator: peripheral patterns require a positive HIT, unidirectional nystagmus, and no skew. Any deviation from that pattern warrants imaging. Other differential considerations:

• Labyrinthitis — same picture plus hearing loss and/or tinnitus on the affected side. Outside the AUVP criteria.²⁸
• Ménière's disease (first attack) — fluctuating unilateral hearing loss is the discriminator. AUVP has stable, preserved hearing.
• Vestibular migraine — episodes usually shorter, with prior migraine history, headache features, and triggers. Spontaneous nystagmus is uncommon between attacks.
• Anterior vestibular artery infarct — produces an identical clinical picture but with central origin. Anterior vestibular artery distribution overlaps the superior division territory; ipsilateral hearing loss and additional central signs argue against AUVP.²⁸
• Selective inferior division neuritis — acute vertigo with normal calorics. Easy to miss if the test battery does not include cVEMP and posterior canal vHIT.^30,31

Reading the report

An asymmetric DVA loss greater on head motion toward the affected ear, combined with documented unilateral VOR loss on horizontal vHIT or caloric, in a patient with a self-limited acute vertigo episode lasting ≥24 hours and no hearing change, is sufficient for the diagnosis under the Bárány 2022 criteria.²⁸

When the asymmetric DVA is the only abnormal vestibular finding — calorics normal, horizontal vHIT normal — pursue the inferior division pattern explicitly: cVEMP, posterior canal vHIT, and saccule-specific symptoms. Failure to do so under-diagnoses up to several percent of neuritis cases that the routine battery would otherwise miss.^30,31

1. 0:00Ménière's disease is the disease in which DVA is the secondary actor, not the lead. The diagnosis is anchored on episodic vertigo plus fluctuating low-frequency sensorineural hearing loss in the affected ear. The audiogram is the pivotal test here. DVA helps characterise the functional consequence of the vestibular failure between attacks and during the chronic phase.
2. 0:30The Bárány Society 2015 criteria are a joint consensus across the Bárány Society, the Japan Society for Equilibrium Research, the European Academy of Otology and Neurotology, the American Academy of Otolaryngology, and the Korean Balance Society. They define two categories: definite and probable Ménière's disease.
3. 1:00Definite Ménière's requires at least two spontaneous vertigo episodes lasting twenty minutes to twelve hours, audiometrically documented low- to mid-frequency sensorineural hearing loss in the affected ear at one moment in time around an episode, fluctuating aural symptoms — hearing, tinnitus, or fullness — and exclusion of other causes. Probable Ménière's allows a broader symptom window of twenty minutes to twenty-four hours with fluctuating aural features but without a strict audiometric requirement.
4. 1:35The DVA picture varies with disease stage. Early on, between attacks, DVA may be entirely normal. After repeated episodes the affected ear accumulates a chronic vestibular deficit; DVA then shows an asymmetric loss that resembles vestibular neuritis but without the acute vertigo and with a documented audiometric fluctuation pattern.
5. 2:10A useful diagnostic signature is the dissociation between the caloric test and the video head-impulse test. In Ménière's the caloric is commonly abnormal on the affected side while the vHIT is normal — a pattern that occurs in roughly half of patients in the published meta-analysis. McGarvie and colleagues attribute this to enlargement of the semicircular canal duct in the hydropic labyrinth, which reduces the caloric thermal gradient but leaves rotational responses intact.
6. 2:50Two practical points. First, the audiogram is the diagnostic test, not the DVA. Document the hearing fluctuation. Second, use DVA to track the chronic vestibular deficit between attacks and to monitor response to intratympanic treatment, where unilateral vestibular loss is the intended therapeutic effect.

What is Ménière's disease?

Ménière's disease is an episodic inner-ear disorder characterised by recurrent attacks of spontaneous vertigo, fluctuating sensorineural hearing loss, tinnitus, and a sense of aural fullness in the affected ear. The histopathological substrate is endolymphatic hydrops — dilatation of the endolymphatic compartment of the inner ear — though the mechanistic link from hydrops to symptoms remains incompletely understood.³²

The disease is unilateral at presentation in approximately 80–90% of cases; bilateral involvement may develop over years. Attacks last from twenty minutes to twelve hours, with prolonged unsteadiness afterwards that may take days to settle. Between attacks the patient may be entirely asymptomatic or carry a residual hearing deficit and chronic vestibular insufficiency.³²

How common is it?

Population prevalence estimates vary widely depending on case definition and ascertainment, ranging from roughly 13 to 190 per 100,000 people. The condition typically presents in the fourth to sixth decade, with a slight female predominance.³²

Bárány Society 2015 diagnostic criteria

The 2015 consensus, jointly endorsed by the Bárány Society, the Japan Society for Equilibrium Research, EAONO, AAO-HNS, and the Korean Balance Society, defines two diagnostic categories.³²

Definite Ménière's disease

1. Two or more spontaneous episodes of vertigo, each lasting 20 minutes to 12 hours.
2. Audiometrically documented low- to mid-frequency sensorineural hearing loss in the affected ear, defining or causing the diagnosis, on at least one occasion before, during, or after one of the episodes of vertigo.
3. Fluctuating aural symptoms (hearing, tinnitus, or fullness) in the affected ear.
4. Not better accounted for by another vestibular diagnosis.

Probable Ménière's disease

A broader category for the early or partial picture: episodic vestibular symptoms (vertigo or dizziness) lasting 20 minutes to 24 hours, with fluctuating aural symptoms, and not better accounted for by another diagnosis. The audiometric requirement is relaxed to allow diagnosis before objective hearing-loss documentation.³²

The caloric–vHIT dissociation

A characteristic and diagnostically useful finding in Ménière's disease is the dissociation between caloric testing and the video head impulse test. Despite both nominally probing the horizontal canal, they often disagree.^33,34

• Caloric is commonly abnormal on the affected side. A meta-analysis of 708 Ménière's patients reported altered caloric responses in approximately 64% of cases.³³
• vHIT is commonly normal in the same patients — altered in only about 28%. The dissociation pattern (abnormal caloric, normal vHIT) appears in roughly 47% of patients.³³

McGarvie and colleagues proposed an anatomical explanation: physical enlargement of the semicircular canal duct in the hydropic labyrinth reduces the thermally-induced pressure gradient across the cupula (impairing the caloric response) while leaving the rotational response (and therefore the vHIT) largely intact.³³ A 2024 confirmatory study in 2,101 patients showed the dissociation to be substantially more common in Ménière's than in vestibular migraine or other vestibular disorders — supporting its use as a differentiating feature.³⁴

The DVA picture by disease stage

DVA is unhelpful as a diagnostic test in Ménière's disease and useful as a longitudinal one. The pattern evolves predictably:

• Inter-ictal, early disease: DVA frequently normal. The patient has full vestibular function between attacks, with the caloric paresis being a low-frequency lab finding rather than a high-frequency clinical deficit. Bedside DVA is therefore insensitive at this stage.
• Per-ictal, during an attack: DVA is uninterpretable — the patient cannot tolerate head motion. Audiometry and the character of the spontaneous nystagmus (irritative pattern beating toward the affected ear early, paretic pattern beating away later) are the clinically useful tests.
• Chronic, advanced disease: Repeated attacks accumulate a chronic vestibular deficit; DVA shows an asymmetric loss greater on the affected side. The pattern resembles compensated vestibular neuritis. It is at this stage that DVA earns its keep in Ménière's — to quantify the residual deficit and to guide vestibular rehabilitation.^3,5
• Post-ablation: Intratympanic gentamicin and labyrinthectomy aim to abolish vestibular function on the affected side. DVA on the affected side falls precipitously, then partially recovers as central compensation deploys covert saccades over weeks to months — exactly the pattern seen after vestibular neuritis. DVA is the right outcome measure for this trajectory.^5,25

Treatment monitoring with DVA

Three Ménière's interventions produce predictable DVA trajectories:

• Intratympanic steroid (dexamethasone) — does not ablate vestibular function. DVA should not change. If it worsens after intratympanic steroid, investigate other causes.
• Intratympanic gentamicin — chemical labyrinthectomy. Affected-side DVA worsens within days to weeks of treatment, then partially recovers through central compensation. Pre-treatment and 6- to 12-week post-treatment DVA on the same paradigm is a useful outcome pair.
• Labyrinthectomy or vestibular nerve section — complete ablation. Affected-side DVA loss is profound and permanent. Rehabilitation focuses on developing covert catch-up saccades. The DVA recovery pattern mirrors the Herdman 2003 unilateral rehabilitation evidence.^3,5

Differential diagnosis

• Vestibular migraine — episodic vertigo with migraine features. Hearing loss is uncommon and rarely documented audiometrically. The Mavrodiev 2024 confirmatory cohort showed the caloric–vHIT dissociation strongly favoured Ménière's over vestibular migraine.³⁴
• Vestibular schwannoma — gradually progressive unilateral SNHL plus episodic vertigo can mimic Ménière's. MRI is the discriminator; consider it for any unilateral SNHL.
• Autoimmune inner-ear disease — rapidly progressive bilateral SNHL with vertigo. Considered when the disease pattern does not fit the Ménière's episodic criteria.
• Perilymph fistula — vertigo and hearing change triggered by pressure or trauma. History is the key.
• Tumarkin "otolithic crisis" / drop attacks — rare, late complication of Ménière's itself. Sudden falls without warning during the active disease phase.

Reading the report

A patient with two or more episodes of vertigo lasting 20 minutes to 12 hours, audiometrically documented low- to mid-frequency SNHL on the affected side, and fluctuating aural symptoms meets the Bárány 2015 definite Ménière's criteria.³² The vestibular test pattern — caloric paresis with normal vHIT (the dissociation) — supports the diagnosis but is not required.^33,34 DVA contributes to the chronic-phase assessment and to monitoring the response to ablative interventions, not to the initial diagnosis.

1. 0:00Vestibular schwannoma is the imaging-anchored disease. The audiogram and the patient's history identify who to image. The MRI is the diagnostic test. DVA and the rest of the vestibular battery serve as supporting evidence and as monitoring tools — they don't make the diagnosis.
2. 0:25The tumour is a benign Schwann-cell neoplasm arising from the vestibular branch of the eighth cranial nerve, most often the inferior vestibular division. It grows slowly — typically about one millimetre per year — within the internal auditory canal and the cerebellopontine angle. Sporadic tumours are unilateral; bilateral tumours raise suspicion of neurofibromatosis type 2.
3. 0:55The classic presentation is asymmetric sensorineural hearing loss, typically high-frequency and slowly progressive, plus unilateral tinnitus and a vague sense of imbalance. Rotational vertigo is uncommon because the slow tumour growth allows continuous central compensation.
4. 1:25The MRI gold standard is gadolinium-enhanced T1 imaging of the internal auditory canal. The audiogram is the screening test that identifies who should have an MRI. Any unilateral or asymmetric sensorineural hearing loss warrants imaging, full stop. Audiometric screening protocols cannot reliably exclude schwannoma — the Bayraktar 2022 meta-analysis and the Reilly 2024 deep-learning study both concluded that no audiometric pattern is sensitive enough to forgo MRI.
5. 2:00The vestibular test battery has variable sensitivity. The West 2020 cohort reported caloric sensitivity of 93%, vHIT 80%, and cVEMP 73%. No single vestibular test is sensitive enough to exclude the diagnosis. DVA correlates with vHIT findings — when vHIT is abnormal, DVA is usually abnormal too — but DVA is also often normal because slow tumour growth allows the central nervous system to compensate.
6. 2:35Where DVA earns its keep here is post-surgically. After translabyrinthine resection, the vestibular function on the affected side is abolished. The patient now has an acute unilateral vestibulopathy on top of an already-compensated chronic lesion. DVA quantifies the deficit, and vestibular rehabilitation tracked by DVA is the standard post-operative care.

What is a vestibular schwannoma?

A vestibular schwannoma is a benign tumour of Schwann cells — the glial cells that myelinate peripheral nerves — arising from the vestibular branch of the eighth cranial nerve. Most arise from the inferior vestibular division. The tumour begins within the internal auditory canal (IAC) and grows medially out into the cerebellopontine angle (CPA) as it enlarges.³⁷

Growth is typically slow — about 1 mm per year on average — which allows the central nervous system continuous opportunity to compensate. Acute rotational vertigo is uncommon at presentation; patients more often describe vague unsteadiness, asymmetric hearing loss, and unilateral tinnitus.³⁷

How common is it?

Detected incidence has risen substantially over recent decades — from roughly 1 per 100,000 person-years in the 1970s to 3–5 per 100,000 person-years today. The increase is largely attributable to MRI availability and incidental detection on imaging performed for unrelated reasons, rather than a true epidemiological rise. The tumour typically presents in the fourth to sixth decade and affects men and women equally.^36,37

Bilateral schwannomas are the hallmark of neurofibromatosis type 2 (NF2), an autosomal-dominant genetic disorder. Five per cent of patients with apparently sporadic schwannomas develop contralateral disease and should be assessed for NF2.³⁷

Who needs MRI?

Any patient with unilateral or asymmetric sensorineural hearing loss, unilateral non-pulsatile tinnitus, or unexplained unilateral vestibular dysfunction warrants gadolinium-enhanced MRI of the internal auditory canal. The 2020 European Association of Neuro-Oncology (EANO) guideline considers this the diagnostic gold standard.³⁷

Multiple audiometric screening protocols have been proposed — "rule of 3000," asymmetry of ≥15 dB at adjacent frequencies, asymmetric speech discrimination, and so on. Systematic review and recent deep-learning analysis show that no audiometric pattern is sensitive enough to reliably exclude schwannoma in a patient where it is clinically suspected.³⁷ The screening question is not whether the audiogram is "abnormal enough" — it is whether the asymmetry is unexplained.

The vestibular test battery

West and colleagues (2020) reported test-battery sensitivity in 59 patients with surgically confirmed unilateral vestibular schwannoma:³⁵

• Caloric test: 93% sensitivity
• vHIT: 80% sensitivity
• cVEMP: 73% sensitivity

Several patterns are worth recognising:

• Caloric weakness with normal vHIT is the single most common pattern — analogous to the Ménière's dissociation but for a different reason. The slow tumour growth allows compensation that masks the rotational deficit at high frequencies even when low-frequency function (caloric) is impaired.³⁵
• cVEMP loss reflects the tumour's preferential arising from the inferior vestibular division.³⁷
• Posterior canal vHIT abnormality with preserved horizontal canal vHIT can be the earliest vestibular sign — again reflecting inferior division origin. Worth looking for if only horizontal vHIT is performed and is normal.

Where DVA fits

DVA is neither sensitive nor specific enough to screen for or to exclude vestibular schwannoma. Older estimates report sensitivity around 81% and specificity around 53% — useful as supporting evidence but never as a primary diagnostic test.^1,35

Three contexts in which DVA earns its keep in this disease:

• Pre-operative baseline. If the patient is heading to surgery, a documented baseline DVA frames how much functional loss the surgery will add and how much rehabilitation will be needed.
• Post-operative monitoring. Translabyrinthine resection ablates vestibular function on the affected side. DVA quantifies the acute deficit, then tracks recovery through covert catch-up saccades over weeks to months. This is the same trajectory seen after intratympanic gentamicin in Ménière's and after vestibular neuritis recovery.^3,5
• Surveillance during observation. Many small tumours are managed with serial MRI rather than surgery. A patient developing oscillopsia or DVA decline during observation may have growth even if MRI changes are subtle. DVA is not a replacement for imaging, but it is a useful symptom-anchored adjunct.

Management overview

The EANO 2020 guideline frames three principal management options, chosen against tumour size, growth rate, age, hearing status, and patient preference:³⁷

• Observation with serial MRI.The default for small (< 15 mm intracanalicular) or incidentally detected tumours with limited symptoms. Most show slow or absent growth.
• Stereotactic radiotherapy / radiosurgery. For small to medium tumours. Aims to halt growth rather than to remove. Hearing preservation rates depend on baseline hearing and tumour size.
• Microsurgical resection. For large tumours, those causing brainstem compression, those failing radiotherapy, or where patient preference favours definitive treatment. Three principal approaches: translabyrinthine (sacrifices hearing; widest exposure for large tumours), retrosigmoid (preserves hearing where feasible), and middle cranial fossa (best for small intracanalicular tumours with serviceable hearing).

Post-surgical DVA: a unilateral vestibulopathy in slow motion

Translabyrinthine resection abolishes vestibular function on the tumour side. The clinical picture immediately after surgery resembles acute vestibular neuritis — spontaneous nystagmus, severe unsteadiness, asymmetric DVA. The trajectory is then identical to neuritis recovery: static compensation removes the spontaneous nystagmus over days to weeks, and structured vestibular rehabilitation drives DVA recovery over weeks to months.^3,5,25

Two practical patterns deserve attention:

• Pre-existing compensation accelerates recovery. A patient with long-standing tumour growth has often already compensated centrally for much of the vestibular asymmetry. After surgery the recovery is therefore faster than after an acute neuritis of similar magnitude — the brain has had years of practice.
• Older patients compensate more slowly and may never fully recover DVA. The Anson 2016 data on age and compensatory saccades is relevant here: older brains generate larger but later catch-up saccades. Plan rehabilitation timing accordingly.¹⁵

Differential diagnosis

• Sudden sensorineural hearing loss (SSNHL) — but schwannoma can rarely present this way too. SSNHL workup should include MRI to exclude tumour.
• Meningioma of the CPA — distinguishable on MRI by wider base on the petrous bone and dural tail. Vestibular tests may be similar.
• Epidermoid cyst of the CPA — distinct MRI signal (high T2, DWI restriction).
• Facial schwannoma in the IAC — rare; facial weakness disproportionate to hearing loss.
• Ménière's disease — episodic vertigo plus low-frequency fluctuating SNHL distinguishes from the high-frequency progressive pattern of schwannoma. Caloric–vHIT dissociation occurs in both; MRI is the discriminator.^32,35

Reading the report

A patient with unilateral or asymmetric sensorineural hearing loss, disproportionately reduced speech discrimination, and an enhancing mass in the IAC or CPA on gadolinium-enhanced MRI has a vestibular schwannoma until proven otherwise.³⁷ Vestibular findings — including DVA — are supporting evidence and matter most for surgical planning and post-operative care. A normal DVA does not exclude the diagnosis; an abnormal DVA does not establish it.³⁵

1. 0:00Vestibular migraine is the most common cause of recurrent spontaneous vertigo in neurology clinics, and probably the most under-recognised. The diagnosis is purely clinical — there is no biological marker, no signature test result, no MRI finding. The Bárány Society and the International Headache Society wrote the criteria together in 2012, with a literature update in 2022. The criteria themselves are unchanged.
2. 0:30Definite vestibular migraine requires four things. First, at least five episodes of moderate or severe vestibular symptoms lasting five minutes to seventy-two hours. Second, a current or past history of migraine with or without aura by ICHD criteria. Third, migraine features — headache, photophobia and phonophobia, or visual aura — in at least half of the vestibular episodes. Fourth, exclusion of other vestibular and headache diagnoses.
3. 1:10Probable vestibular migraine relaxes the third criterion: the episodes meet the duration and intensity requirements, but only one of the migraine criteria is met — either the history of migraine, or migraine features in the episodes, but not both.
4. 1:40Vestibular migraine differs fundamentally from every other disease in this atlas. It is a central disorder of visuo-vestibular integration rather than a peripheral end-organ disease. Vestibular testing in vestibular migraine is heterogeneous: roughly half of patients have mild abnormalities, half have none, and the pattern rarely localises the way it does in neuritis or Ménière's. The vHIT is usually normal even when the patient is severely symptomatic.
5. 2:15DVA in vestibular migraine reflects the central pathophysiology. It can be abnormal in all four directions — left, right, up, and down — rather than the directional asymmetry seen in unilateral peripheral disease. The pattern is symmetric but the magnitude is typically modest. This is the head-motion intolerance and visual motion sensitivity that migraine patients describe between attacks, made measurable.
6. 2:45Two practical points. First, vestibular migraine and Ménière's disease overlap clinically in the first year — the patient with episodic vertigo and aural symptoms may meet either diagnosis. The audiogram and the Mavrodiev 2024 caloric–vHIT dissociation help distinguish, but time and serial assessment matter. Second, vestibular migraine often coexists with anxiety and persistent postural-perceptual dizziness. Recognise these overlaps — treating only the vestibular component misses half the picture.

What is vestibular migraine?

Vestibular migraine (VM) is the second most common cause of recurrent vertigo after benign paroxysmal positional vertigo, and the most common cause of spontaneous recurrent vertigo in neurology clinics. It affects approximately 1% of the general population and accounts for about 11% of patients in tertiary dizziness clinics.³⁸

Unlike every other disease in this atlas, vestibular migraine is not an end-organ disorder. The vestibular system itself is structurally intact. The pathophysiology is central — abnormal processing of vestibular and visual motion signals in the brain, related mechanistically to migraine itself. This is why the diagnosis is purely clinical, and why peripheral vestibular tests (vHIT, calorics, VEMPs, even most DVA assessments) are typically normal or only mildly abnormal.^38,39

Bárány Society / IHS 2012 criteria

The criteria were jointly formulated by the Bárány Society's Classification Committee and the International Headache Society's Migraine Classification Subcommittee in 2012, with a literature update in 2022. The criteria themselves are unchanged between the two documents. Vestibular migraine is recognised in the appendix of ICHD-3 as a condition needing further research.^38,39

Definite vestibular migraine

All four of:

1. At least five episodes of moderate or severe vestibular symptoms (vertigo, spontaneous or positional or head-motion-induced; or head-motion-induced dizziness with nausea), lasting 5 minutes to 72 hours.
2. Current or previous history of migraine with or without aura, per ICHD criteria.
3. One or more migraine features accompanying at least 50% of the vestibular episodes:
   • headache with at least two of: unilateral location, pulsating quality, moderate/severe intensity, aggravation by routine activity
   • photophobia and phonophobia
   • visual aura
4. Not better accounted for by another vestibular or ICHD diagnosis.

Probable vestibular migraine

Episodes meeting criterion 1 (≥5 episodes, 5 min–72 h, moderate or severe), only one of criteria 2 and 3 (migraine history or migraine features per episode but not both), and exclusion of other diagnoses.

The duration distribution

Episode duration in vestibular migraine is famously variable. The consensus document reports:

• ~30% of patients: episodes lasting minutes
• ~30% of patients: episodes lasting hours
• ~30% of patients: episodes lasting several days
• ~10% of patients: very short attacks of seconds, triggered by head motion, visual stimulation, or positional change. The "episode" in this group is defined as the total period during which short attacks recur.³⁸

This spread alone places vestibular migraine in the differential for nearly any episodic vertigo pattern other than the truly fixed windows (BPPV under 60 seconds; Ménière's 20 minutes to 12 hours).

The vestibular test battery — heterogeneous and unhelpful

Inter-ictal vestibular testing in VM is heterogeneous. About half of patients show mild abnormalities, but the pattern rarely localises. The 2012 consensus is explicit: "Vestibular findings and testing results can be pathological, particularly during or shortly after an episode, but they are not sufficiently specific to serve as diagnostic criteria."³⁸

Recent ictal-versus-interictal work (2025) found that VOR function on vHIT remained normal across attack phases, while subjective visual vertical deviation was significantly worse ictally — consistent with central rather than peripheral dysfunction.

Profound abnormalities — bilateral vestibular loss, severe unilateral SNHL, complete unilateral vestibular loss — should prompt re-evaluation. They are not part of the VM picture, and they suggest another diagnosis or a comorbid one.³⁸

The DVA pattern in detail

DVA findings in vestibular migraine reflect the central pathophysiology. Three patterns are worth recognising:

• Symmetric mild loss in all four directions — left, right, up, down. Sevimli and colleagues (2022) showed significant DVA loss in all four positions in twenty migraine patients compared with controls, interpreted as evidence of impaired visuo-vestibular cortical integration rather than peripheral VOR failure.⁴⁰ The shape of the pattern resembles bilateral vestibulopathy (no asymmetry), but the magnitude is much smaller (≤0.2 logMAR) and the underlying mechanism is entirely different.
• Ictal worsening with inter-ictal recovery. DVA measured during an attack is worse than between attacks — but most testing is done in the inter-ictal phase, where it appears modestly or not at all abnormal.
• Visual motion sensitivity correlate. The DVA asymmetry between active and passive head motion paradigms is sometimes wider in VM than in controls, fitting with the head-motion intolerance these patients describe. This is more a research finding than a clinical one at present.

Distinguishing VM from Ménière's disease

Vestibular migraine and Ménière's disease overlap clinically in the first year of symptoms, and the two diseases co-occur more often than chance — migraine is more common in Ménière's patients than in healthy controls, and bidirectional inheritance has been described.^32,38 Distinguishing the two matters because management differs.

Useful discriminators:

• Audiogram. Definite Ménière's requires audiometrically documented low- to mid-frequency SNHL on the affected side. VM does not produce sustained SNHL — fluctuating hearing in VM tends to be transient and recovers fully.^32,38
• Caloric–vHIT pattern. The Mavrodiev 2024 study in 2,101 patients established that the dissociation pattern (abnormal caloric, normal vHIT) is far more common in Ménière's than in VM. In a patient where the diagnosis is uncertain, this pattern argues for Ménière's.³⁴
• Migraine features in episodes. Photophobia and phonophobia during attacks favour VM. Auditory symptoms (fullness, tinnitus, distortion) localised to one ear favour Ménière's.
• Episode duration distribution. Ménière's episodes cluster at hours; VM episodes spread from minutes to days. The very-short and the very-long ends of the duration spectrum favour VM.³⁸

When the picture is genuinely mixed and both criteria are met, the Bárány Society allows diagnosis of both conditions if the patient has two distinguishable types of episodes. A future revision may include an overlap syndrome category.³⁸

Other differential diagnoses

• BPPV. Vestibular migraine with positional features can mimic BPPV. In VM, the positional nystagmus is usually persistent (not paroxysmal) and does not align with a single canal; symptomatic episodes are shorter and more frequent than in BPPV.³⁸
• Persistent postural-perceptual dizziness (PPPD). The chronic between-attack dizziness in VM can evolve into PPPD — the two coexist in many patients. PPPD is characterised by persistent (≥3 months) dizziness exacerbated by upright posture, self-motion, and complex visual stimuli.
• Vestibular paroxysmia. Brief (seconds) attacks of vertigo recurring many times per day, typically responsive to carbamazepine. The 10% of VM patients with attacks lasting seconds are the differential challenge here.
• Transient ischaemic attacks. Sudden onset, vascular risk factors, total history of attacks less than one year, and evidence of vertebrobasilar pathology favour TIA over VM, especially in older patients.³⁸
• Anxiety-related dizziness. Common comorbidity with VM (over 50% in some series). Situational provocation, autonomic activation, and avoidance behaviour are clues. Treat both.³⁸

Management implications for the test request

When VM is the leading diagnosis, the role of vestibular testing shifts from establishing the diagnosis to excluding alternatives. Three orders worth considering:

• MRI of the brain and IAC — to exclude structural disease (schwannoma, demyelination, posterior fossa lesion). VM does not have an MRI signature.
• Audiometry — to exclude Ménière's disease and to document the baseline. Repeat during a symptomatic period if the initial audiogram is normal.
• vHIT and caloric — to exclude unilateral or bilateral peripheral vestibulopathy. The classic "VM" result is both being broadly normal.

Reading the report

A patient with ≥5 episodes of moderate-to-severe vertigo lasting 5 minutes to 72 hours, a migraine history, migraine features in at least half of the episodes, and exclusion of alternative diagnoses, meets the Bárány 2012 definite criteria for vestibular migraine.³⁸ DVA, vHIT, caloric, and audiometry contribute by what they exclude, not by what they confirm. A broadly normal vestibular test battery in the right clinical context supports the diagnosis precisely because it makes alternatives less likely.^38,39

1. 0:00Central causes are the diseases this atlas spends every prior chapter trying to exclude. They are also the diseases that DVA alone cannot diagnose — and the ones where missing the diagnosis has the highest stakes. The most clinically important question in vestibular medicine is not 'what kind of peripheral vestibular disease does this patient have' but 'is this patient peripheral at all'. This chapter is about that question.
2. 0:35The clinical setting that matters most is the acute vestibular syndrome — sudden onset of vertigo, nausea, head-motion intolerance, and unsteadiness, lasting at least 24 hours. Most patients with this presentation have vestibular neuritis. A minority have posterior circulation stroke. The difference matters enormously — neuritis is benign and self-limiting; stroke kills people, particularly when the cerebellar oedema progresses over the first 48 hours.
3. 1:10The HINTS exam — Head-Impulse, Nystagmus, Test-of-Skew — is the bedside discriminator. Kattah and Newman-Toker showed in 2009 that in high-risk acute vestibular syndrome patients, a dangerous HINTS result was 100% sensitive and 96% specific for central lesions, outperforming early MRI diffusion-weighted imaging. The dangerous pattern is the INFARCT mnemonic: Impulse Normal, Fast-phase Alternating, Refixation on cover test.
4. 1:50The most counterintuitive piece for trainees is the head-impulse test. A normal HIT in a patient with acute vertigo is <em>dangerous</em>, because it suggests the vestibular nerve is intact and the lesion must be central. An abnormal HIT with catch-up saccades is <em>reassuring</em>, because it localises the lesion to the periphery. This reversal trips up clinicians who learned HIT first in the context of unilateral peripheral disease.
5. 2:25The vestibular test battery in established central disease is heterogeneous and not localising in itself. DVA may be abnormal but the pattern does not localise the lesion. vHIT is often normal because the peripheral vestibular system is intact. Calorics are usually normal. The diagnosis is made on bedside oculomotor findings, additional central neurological signs, and MRI — not on the vestibular test battery.
6. 3:00Three practical points. First, in any patient with acute vestibular syndrome and vascular risk factors, the HINTS exam is mandatory before any peripheral diagnosis is made. Second, MRI within 48 hours of onset misses up to 20% of posterior circulation strokes — the false-negative rate is much higher than for anterior circulation strokes. Third, DVA's role in central disease is to characterise the functional consequence and to guide rehabilitation; it does not contribute to the central-versus-peripheral question, for which HINTS plus additional central signs is the standard.

Why central causes matter

Every previous disease page in this atlas covered a peripheral vestibular disorder — an end-organ disease or eighth-nerve disease in which the vestibular system itself is damaged. Central causes are different in mechanism and different in stakes. The vestibular periphery is intact; the lesion is in the brainstem vestibular nuclei, the cerebellum, the medial longitudinal fasciculus, the thalamic relays, or the cortical vestibular network. The most common and most consequential cause is posterior circulation stroke.⁴⁸

The clinical setting that matters most is the acute vestibular syndrome (AVS) — sudden onset of vertigo, nausea, head-motion intolerance, and unsteadiness lasting ≥ 24 hours. Most AVS is vestibular neuritis. A clinically important minority — around 10–25% in high-risk populations — is posterior circulation stroke. Neuritis is benign and self-limiting; stroke can kill the patient through cerebellar oedema, herniation, and brainstem compression over the first 48 hours.^46,48

The HINTS exam

The Head-Impulse, Nystagmus, Test-of-Skew (HINTS) exam was formalised by Kattah and Newman-Toker in 2009 from prior bedside oculomotor work. In a prospective study of 101 high-risk AVS patients, the dangerous-HINTS battery was 100% sensitive and 96% specific for central lesions — and notably outperformed early MRI diffusion-weighted imaging, which has a significant false-negative rate in posterior circulation stroke within the first 48 hours.⁴⁶

The "dangerous" pattern is captured by the INFARCT mnemonic: Impulse Normal, FAst-phase Alternating, Refixation on Cover Test. Any one dangerous feature in an AVS patient is sufficient to raise concern for stroke and warrant urgent neuroimaging.⁴⁶

Additional bedside central signs

HINTS is highly sensitive but not the only set of useful findings. The Newman-Toker neuro-vestibular examination identifies several additional central oculomotor signs worth checking in any AVS patient — particularly those with vascular risk factors:

• Impaired smooth pursuit — saccadic intrusions during slow eye-tracking; floccular or parafloccular dysfunction.
• Impaired VOR cancellation — patient cannot suppress VOR when head and target move together; the bedside sign least attributable to peripheral disease.
• Gaze-evoked nystagmus — beats in the direction of eccentric gaze; failure of the neural integrator. (Distinct from direction-changing nystagmus on the HINTS exam, which is a related but specific finding.)
• Vertical nystagmus (downbeat or upbeat in primary position) — almost always central.
• Severe truncal instability — inability to sit unaided is suggestive of cerebellar disease.
• Limb dysmetria, dysdiadochokinesis, dysarthria — classical cerebellar signs, but frequently absent in inferior cerebellar infarcts.
• Other focal neurological signs — internuclear ophthalmoplegia, gaze palsy, hemiparesis, sensory loss, facial weakness, dysphagia. Present in approximately 42% of central AVS cases in prospective studies.⁴⁶

Vascular causes — Bárány 2022

The 2022 Bárány Society consensus on vascular vertigo and dizziness classifies vascular causes by symptom duration and mechanism:⁴⁸

• Acute prolonged vascular vertigo/dizziness — ≥ 24 h. Posterior circulation stroke (PICA, AICA, vertebrobasilar territory) or, rarely, isolated labyrinthine infarction.
• Transient vascular vertigo/dizziness— minutes to < 24 h. Posterior circulation TIA. Recurrent transient vestibular symptoms preceding stroke are an important warning sign that should not be dismissed as benign.
• Isolated labyrinthine infarction — vascular disease that does involve the inner ear via the labyrinthine artery (terminal AICA branch). The picture is identical to vestibular neuritis with cochlear involvement. May precede ponto-cerebellar infarction.
• Vertebral artery compression syndrome — vertigo on sustained eccentric neck position with imaging evidence of posterior-circulation flow reduction.

Non-vascular central causes

• Demyelinating disease (MS) — episodic vertigo from brainstem demyelination. May produce INO, gaze-evoked nystagmus, dysmetric saccades. Vertigo can be the presenting symptom in 5–15% of MS cases.
• Cerebellar degeneration — including spinocerebellar ataxias and the CANVAS syndrome (cerebellar ataxia + sensory neuronopathy + vestibular areflexia). Slowly progressive rather than acute.²⁶
• Brainstem tumours — gliomas, metastases, cavernomas. Subacute and progressive.
• Wernicke encephalopathy — thiamine deficiency producing the classic triad of confusion, ataxia, and oculomotor findings (often bilateral abducens palsy with nystagmus). Reversible if treated promptly.
• Migrainous brainstem aura — see also the Vestibular Migraine page. Distinct from vestibular migraine; the ICHD criteria require additional posterior-circulation aura features.
• Superficial siderosis — chronic CSF haemosiderin deposition; slowly progressive cerebellar and vestibular symptoms.

MRI caveats in acute AVS

A common clinician error is to over-rely on MRI in acute AVS. Two quantitative limits matter:

• False-negative rate within 48 hours: Posterior circulation stroke has a much higher MRI false-negative rate than anterior circulation stroke — up to 20% of small cerebellar or brainstem infarcts are missed on initial diffusion-weighted imaging in the first 24–48 hours. The lesion may appear only on repeat imaging at 48–72 hours.^46,48
• HINTS outperforms early MRI: In the Kattah 2009 study, the dangerous-HINTS battery was 100% sensitive for central lesions, while early DWI had a false-negative rate that allowed several strokes to be missed. The bedside exam should not be short-circuited by a "negative MRI" — repeat imaging and clinical observation are warranted when HINTS is dangerous.⁴⁶

Cerebellar stroke syndromes worth knowing

• PICA (posterior inferior cerebellar artery) stroke — most common cerebellar-stroke cause of isolated vertigo. Affects inferior cerebellum and lateral medulla (Wallenberg syndrome). Bedside HIT typically normal; direction-changing nystagmus often present. May appear "isolated" on examination — this is the most dangerous mimic of vestibular neuritis.⁴⁷
• AICA (anterior inferior cerebellar artery) stroke — distinctive because it can involve the labyrinthine artery and produce ipsilateral peripheral signs (abnormal HIT, hearing loss, peripheral nystagmus pattern) alongside central signs. The HINTS exam may be misleadingly "negative" — additional features such as ipsilateral facial weakness, sensory change, or ataxia are crucial.⁴⁸
• SCA (superior cerebellar artery) stroke — usually produces obvious cerebellar signs (limb ataxia, dysarthria) rather than isolated vertigo. Less likely to be confused with vestibular neuritis.
• Vertebrobasilar TIA — recurrent transient vestibular symptoms in a patient with vascular risk factors should be investigated as TIA until proven otherwise. The Kim 2021 study (cited in the Bárány vascular criteria) showed that transient vestibular symptoms commonly precede posterior circulation stroke.⁴⁸

What DVA contributes in central disease

DVA does not contribute to the central-versus-peripheral discrimination at the acute presentation. Its role in established central disease is similar to its role in established Ménière's or schwannoma — quantifying the functional consequence and guiding rehabilitation.

• Post-stroke rehabilitation: Cerebellar stroke survivors with persistent oscillopsia and gait instability benefit from vestibular rehabilitation; DVA tracks recovery similarly to peripheral disease, though the underlying mechanism (development of compensatory eye-movement strategies despite intact peripheral VOR) is different. Functional gains are smaller than in pure peripheral disease.
• MS monitoring: DVA can document the impact of relapse and the response to rehabilitation. Useful as a functional outcome measure alongside the Expanded Disability Status Scale and the timed 25-foot walk.
• Cerebellar degeneration: DVA worsens with disease progression. Serial DVA may detect functional decline before clinical examination reveals it.

Reading the report

The central-versus-peripheral question in acute vestibular syndrome is answered by the HINTS exam plus additional central neurological signs — not by the vestibular test battery or by DVA.^46,47 The Bárány 2022 vascular vertigo criteria require acute central signs, dangerous HINTS findings, or vascular risk factors with additional features to support the diagnosis.⁴⁸ DVA's contribution is to the management phase: quantifying functional consequence, guiding rehabilitation, and monitoring recovery or progression. In acute AVS, the message of this chapter is the inverse of every previous chapter: do not let a normal DVA reassure you that the patient does not have a stroke.

1. 0:00Presbyvestibulopathy is the milder, age-related version of bilateral vestibulopathy. It pairs with bilateral disease the way presbycusis pairs with profound deafness, or presbyopia pairs with blindness — a gradient of severity within the same conceptual category. The Bárány Society codified it in 2019, recognising that vestibular function declines with age in a way that is clinically meaningful even short of meeting bilateral vestibulopathy criteria.
2. 0:35The diagnostic criteria require three things. First, the patient must be sixty years of age or older. Second, they must have at least one chronic symptom of unsteadiness, gait disturbance, or recurrent falls lasting three months or longer. Third — and this is the conceptual pivot — vestibular testing must show mild bilateral reduction in VOR function that lies between normal values and the bilateral vestibulopathy thresholds.
3. 1:10The quantitative thresholds map directly onto the bilateral vestibulopathy criteria. The vHIT gain on both sides must be less than 0.8 and greater than 0.6 — the upper bound is below normal, the lower bound is above the bilateral vestibulopathy threshold. The caloric sum on each side must be less than twenty-five degrees per second and greater than six. Rotational chair gain must be greater than 0.1 and less than 0.3 at 0.1 hertz.
4. 1:50The DVA picture mirrors this gradient. Where bilateral vestibulopathy produces severe symmetric loss often greater than 0.4 logMAR, presbyvestibulopathy produces mild symmetric loss typically in the 0.1 to 0.3 logMAR range. The asymmetry pattern is identical to bilateral disease — symmetric, no directional preference — because the underlying anatomy is the same: bilateral hair-cell and afferent decline. The magnitude is what differs.
5. 2:25Crucially, presbyvestibulopathy never exists in isolation. The Bárány criteria explicitly note that PVP typically occurs alongside other age-related sensory declines — presbyopia, presbycusis, peripheral neuropathy of the feet, and cerebellar or extrapyramidal motor change. The clinical phenotype emerges when multiple sensory systems decline together, each one removing a redundancy that previously masked the others.
6. 3:00Two practical points. First, the diagnosis is useful precisely because it makes age-related vestibular decline a legitimate clinical entity rather than a diagnosis of exclusion. Second, DVA-tracked vestibular rehabilitation in this population works the same way as in bilateral vestibulopathy — through development of covert catch-up saccades — and produces clinically meaningful fall-risk reduction even when peripheral function does not recover.

What is presbyvestibulopathy?

Presbyvestibulopathy (PVP) is a chronic vestibular syndrome of unsteadiness, gait disturbance, and recurrent falls in older adults, attributable to mild bilateral age-related decline of the vestibulo-ocular reflex. The Bárány Society codified the diagnosis in 2019 to recognise an entity that had long been described informally as "age-related vestibular loss."⁴¹

The conceptual frame is presbycusis or presbyopia for the vestibular system. Just as age-related hearing loss is not deafness and age-related visual loss is not blindness, age-related vestibular loss is not bilateral vestibulopathy — it is the milder, incomplete form along the same biological continuum. Hair cells, primary vestibular neurons, and central vestibular pathways all show age-related decline. The clinical syndrome emerges when the decline crosses a threshold of functional consequence.⁴¹

How common is it?

Vestibular dysfunction prevalence rises steeply with age. The Agrawal 2009 analysis of the US National Health and Nutrition Examination Survey (n = 5,086) found that around 50% of adults aged 60–69 and approximately 85% of those aged 80 or above had some measurable vestibular dysfunction.⁴² Not all of these patients meet PVP criteria — the diagnosis requires both the quantitative thresholds and a clinical syndrome — but the epidemiological scale is large enough that PVP is one of the most prevalent vestibular diagnoses in the over-60 population.

Bárány Society 2019 diagnostic criteria

The diagnosis of PVP requires all of the following:⁴¹

1. Age ≥ 60 years.
2. At least one of: postural imbalance or unsteadiness, gait disturbance, chronic dizziness, or recurrent falls — lasting ≥ 3 months.
3. Mild bilateral peripheral vestibular deficit documented by VOR testing — i.e. test values that lie between normal and the bilateral vestibulopathy thresholds.
4. Not better accounted for by another disease or disorder.

The threshold gradient

The quantitative thresholds are the conceptual core of the diagnosis. They are deliberately positioned between normal values and the bilateral vestibulopathy thresholds, codifying PVP as the intermediate zone of a continuum rather than a separate disease.^24,41

Meeting any one of the three frequency-range thresholds is sufficient. The three tests probe different parts of the VOR frequency response — vHIT for the high-frequency end, rotational chair for the middle range, calorics for the low-frequency end — and age-related decline can be uneven across the spectrum.⁴¹

The multisensory context

PVP rarely exists in isolation. The Bárány 2019 consensus is explicit that the syndrome typically occurs alongside other age-related sensory declines:⁴¹

• Presbyopia — reduces the visual contribution to postural control, removing a redundancy that previously masked the vestibular decline.
• Presbycusis — age-related sensorineural hearing loss, almost universal in the over-70 population, and often co-occurring with vestibular decline due to shared hair-cell mechanisms.
• Peripheral neuropathy of the feet — reduces somatosensory input from the lower limbs, which the brain otherwise uses to compensate for vestibular asymmetry, particularly in darkness or on uneven ground.
• Mild cerebellar or extrapyramidal change — reduces central capacity to integrate the remaining vestibular, visual, and proprioceptive inputs.

This multisensory context is not coincidence — it is part of why PVP becomes symptomatic when it does. The vestibular system has large reserve. The patient becomes symptomatic when concurrent decline of other systems removes the redundancy.

DVA in PVP vs bilateral vestibulopathy

The shape of the DVA finding is identical between the two diseases — symmetric loss, no directional asymmetry, both directions equally affected. The diagnostic difference is in the magnitude:

• Bilateral vestibulopathy: typical DVA loss ≥ 0.4 logMAR (four chart lines), often higher; absolute loss against age-matched norms is severely below threshold; bedside often shows ≥ four lines lost in both directions.
• Presbyvestibulopathy: typical DVA loss 0.1–0.3 logMAR (one to three lines); absolute loss against age-matched norms is mild-to-moderate; bedside typically shows one to three lines lost in both directions.

For practical interpretation, the question is not "is DVA abnormal?" but "does the DVA loss match the laboratory VOR loss?" A patient with vHIT gain 0.7 bilaterally and DVA loss of 0.2 logMAR has a consistent picture. A patient with vHIT gain 0.7 bilaterally but DVA loss of 0.6 logMAR has either a central contribution (cerebellar disease, visuo-vestibular integration failure) or a measurement artefact — investigate.^1,41

Rehabilitation: less impressive numbers, comparable function gains

Vestibular rehabilitation in PVP works the same way as in bilateral vestibulopathy — development of well-timed covert catch-up saccades during head motion, without a measurable change in peripheral VOR gain.²⁵ The absolute logMAR change with rehabilitation tends to be smaller in PVP than in BVP because the baseline is closer to normal — there is less room to improve in purely numerical terms. The clinically meaningful change is in functional metrics: falls per year, gait speed, dizziness handicap inventory scores. DVA improvement of 0.1 logMAR over six weeks is a typical responder pattern.^5,25

The fall-prevention argument is what makes the diagnosis useful in practice. PVP is a treatable contributor to fall risk in older adults — and falls are a leading cause of morbidity and mortality in this population. Naming the diagnosis legitimises referral to structured rehabilitation and to multidisciplinary fall-prevention services.^41,42

Differential diagnosis

• Bilateral vestibulopathy — same picture, more severe. The thresholds make the distinction quantitative.
• Cerebellar ageing or degeneration — produces similar unsteadiness but with central oculomotor signs (gaze-evoked nystagmus, dysmetric saccades, impaired pursuit). DVA may be abnormal centrally even with preserved vHIT.
• Multisensory unsteadiness of the elderly — the patient has normal individual sensory tests but symptomatic unsteadiness from cumulative mild deficits across vision, proprioception, and vestibular. The PVP diagnosis specifically requires documented vestibular impairment.⁴¹
• Orthostatic hypotension — postural lightheadedness on standing, not present at rest, and not provoked by head motion. Check supine and standing blood pressures.
• Medication-induced unsteadiness — common offenders include antihypertensives, sedatives, anticonvulsants, and aminoglycosides. Review the prescription list explicitly.
• Spinal cord pathology (cervical myelopathy) — produces gait ataxia with hyperreflexia and Babinski sign; vestibular tests are normal.

Reading the report

A patient ≥ 60 years old with ≥ 3 months of unsteadiness, gait disturbance, or recurrent falls, in whom vestibular testing shows bilateral VOR reduction in the PVP threshold band, meets the Bárány 2019 criteria.⁴¹ The DVA finding — symmetric mild loss without directional asymmetry — corroborates the laboratory picture and provides the right outcome measure for rehabilitation. The diagnosis legitimises age-related vestibular decline as a clinical entity and opens the management pathway to structured rehabilitation and fall prevention.^25,41

1. 0:00Ototoxicity is the most preventable cause of bilateral vestibulopathy. The mechanism is identical — bilateral, symmetric, hair-cell-level VOR loss — but the cause is iatrogenic, dose-related, and visible if you look for it. Aminoglycoside antibiotics are the principal offenders. Platinum-based chemotherapy is a growing second source. DVA is one of the surveillance tools that catches the vestibular deficit before it becomes catastrophic.
2. 0:35Aminoglycosides cross into the inner ear and concentrate in vestibular hair cells through the mechanotransduction channels. Type I hair cells in the central crista of each semicircular canal are most vulnerable. Type II cells and the otolith organs are relatively spared. The damage starts within days of exposure, accumulates with cumulative dose, and is usually permanent — hair cells do not regenerate in humans.
3. 1:15Among the aminoglycosides, gentamicin is the most vestibulotoxic agent in widest clinical use, followed by streptomycin and tobramycin. Amikacin and kanamycin are preferentially cochleotoxic rather than vestibulotoxic. This matters because audiometric monitoring alone — which detects cochleotoxicity — completely misses gentamicin vestibulotoxicity. About 90% of patients with gentamicin vestibulotoxicity have no measurable hearing loss.
4. 1:55Clinically the presentation is identical to bilateral vestibulopathy from any other cause. Symmetric oscillopsia with head motion, ataxia worsening in darkness and on uneven ground, no rotational vertigo because both sides are equally affected, no spontaneous nystagmus. The DVA pattern is the same: symmetric loss often greater than 0.4 logMAR, no directional asymmetry, profoundly impaired bedside head-impulse testing.
5. 2:30The surveillance argument is what makes this chapter different. If you screen DVA and bedside head-impulse testing at baseline and at weekly intervals during treatment, you can stop the drug before the damage becomes complete. The dynamic-illegible-E test takes two minutes at the bedside, requires only a Snellen chart, and has been recommended in the Canadian Medical Association Journal practice review for hospital surveillance of patients on aminoglycoside therapy. Most institutions do not currently do this, which is why this chapter exists.
6. 3:05Three practical points. First, treatment duration over seven days, renal impairment, and elevated trough levels are the major risk factors — but vestibulotoxicity occurs at normal trough levels and normal renal function in some patients. Surveillance is not optional for at-risk patients. Second, once detected, stopping the drug is the only effective intervention; vestibular function rarely recovers. Third, DVA-tracked rehabilitation works the same way as in other bilateral vestibulopathies — central compensation through covert saccades. Refer early.

What is ototoxicity?

Ototoxicity is drug- or chemical-induced damage to the inner ear, affecting the cochlea (cochleotoxicity, causing hearing loss), the vestibular system (vestibulotoxicity, causing imbalance and oscillopsia), or both. Aminoglycoside antibiotics are the principal cause of vestibulotoxicity in clinical practice; platinum-based chemotherapy is the second major class. Loop diuretics, salicylates, quinine, and several other agents can contribute, usually reversibly.^43,45

The clinical picture of established aminoglycoside vestibulotoxicity is indistinguishable from bilateral vestibulopathy of any other cause — and indeed, aminoglycoside ototoxicity is the most commonly identified single cause of bilateral vestibulopathy in clinical practice.^26,44 What makes ototoxicity worth its own chapter is the surveillance argument: the damage develops over days to weeks, and if anyone is looking, drug discontinuation during the surveillance window can prevent the catastrophe.^43,45

Agents and mechanism

Aminoglycosides differ in their preferential target within the inner ear:⁴³

• Predominantly vestibulotoxic: gentamicin, streptomycin, tobramycin, netilmicin. Gentamicin is the most vestibulotoxic agent in widest clinical use.
• Predominantly cochleotoxic: amikacin, kanamycin, neomycin, dihydrostreptomycin. These cause hearing loss with relative vestibular sparing.

The mechanism is well characterised. Aminoglycosides enter vestibular hair cells via the mechanotransduction channels at the stereocilia tips — the same channels used in normal vestibular signalling. Once inside, the drug accumulates and damages the cell through oxidative stress, mitochondrial dysfunction, and apoptosis. Type I hair cells in the central crista of each semicircular canal are the first and most severely affected; type II cells and the otolith organs are relatively spared. Damage is dose-dependent and cumulative; once established, it is essentially permanent because human vestibular hair cells do not meaningfully regenerate.⁴³

Platinum chemotherapy

Cisplatin and related platinum-based chemotherapeutic agents have a well-established cochleotoxic profile (high-frequency sensorineural hearing loss in a high fraction of patients) and a less well-characterised but increasingly recognised vestibulotoxic profile. Animal studies show preferential utricular hair-cell loss with cisplatin, contrasting with the canal-crista predilection of aminoglycosides. Human data are heterogeneous, with abnormal vestibular testing reported in 0–50% of patients depending on assessment protocol and cumulative dose; pre-existing vestibular loss increases the risk of further vestibulotoxicity from cisplatin.

Risk factors

• Treatment duration ≥ 7 days is a key threshold. Beyond this, vestibulotoxicity risk rises substantially.⁴⁴
• Renal impairment — slows drug clearance, raises steady-state levels.
• Elevated trough drug levels — though vestibulotoxicity can develop at normal trough levels too.
• Older age, pre-existing hearing loss,concurrent loop diuretic use, previous aminoglycoside exposure.
• Genetic susceptibility— the mitochondrial m.1555A>G variant predisposes to aminoglycoside-induced hearing loss; vestibular equivalents are less well-characterised but similar mechanisms are suspected.

The surveillance workflow

The clinical case for surveillance is straightforward: drug discontinuation during the early surveillance window can limit the final deficit, while detection only after symptom onset usually means established bilateral vestibulopathy. The Canadian Medical Association Journal practice review and the 2018 vestibulotoxicity strategies paper both recommend bedside DVA as a simple, institutional surveillance tool requiring only a Snellen chart and two minutes of clinical time.⁴⁵

A pragmatic surveillance protocol:

1. Baseline (before first dose): bedside DVA, bedside head-impulse test, pure-tone audiogram. Record absolute values, not just "normal."
2. During treatment, days 5 onwards: repeat bedside DVA every 3–4 days. Flag any deterioration of ≥0.2 logMAR (two chart lines) from baseline as a surveillance threshold event.
3. Threshold event triggers: formal vHIT and caloric testing if available, discussion with prescribing team about drug substitution or dose adjustment, and ENT/vestibular referral.
4. Post-treatment (≥6 weeks after final dose): repeat DVA, vHIT, caloric, audiogram for documentation of final deficit and rehabilitation planning.

The DVA pattern in detail

Established aminoglycoside vestibulotoxicity produces the same DVA signature as bilateral vestibulopathy of any other cause — symmetric loss without directional asymmetry, profound impairment of the bedside head-impulse test, near-normal subjective visual vertical. The distinctive feature is not the pattern but the trajectory:

• Days 0–4: DVA typically unchanged from baseline. Hair-cell uptake of the drug is occurring but functional consequence is below the bedside threshold.
• Days 5–14 — the surveillance window: DVA begins to deteriorate, typically asymmetrically at first as the cristae become differentially affected. Loss accelerates around day 7–10. Drug discontinuation in this window limits the final deficit substantially.⁴³
• Days 14–28: the pattern becomes profoundly bilateral. Patient symptoms — oscillopsia, unsteadiness — typically emerge in this window if surveillance has not detected the problem earlier. By this point much of the damage is done.
• Beyond day 28: the deficit plateaus and becomes essentially permanent. Drug cessation prevents further accumulation but does not reverse what has already happened.

What head-impulse saccades look like in vestibulotoxicity

A practical observation worth knowing: bedside head-impulse testing in established aminoglycoside vestibulotoxicity produces unusually large overt saccades — substantially larger than in normal subjects or even in some unilateral vestibulopathies. The cumulative amplitude of overt catch-up saccades has been measured at approximately 5.6 times greater than in healthy controls. Covert saccades, which can mask the bilateral deficit during vHIT interpretation, are only about half as common as in unilateral patients because there is no remaining functional side to provide predictable input.⁴³ Practically, this means the bedside head-impulse test in this population is one of the easier peripheral vestibular signs to observe — the saccades are large and unmissable.

Management once detected

• Stop the offending agent if at all clinically feasible. Discuss with the prescribing team — many infections can be managed with a non-aminoglycoside regimen with comparable efficacy.^43,45
• Refer to vestibular rehabilitation early. DVA-tracked gaze-stabilisation exercises drive the same compensatory-saccade development as in bilateral vestibulopathy of any aetiology — the Herdman 2007 evidence applies directly.²⁵
• Document the deficit formally with vHIT and calorics post-treatment, both for clinical management and because medico-legal documentation can be relevant.⁴⁴
• Address occupational and fall-risk implications. Driving, working at height, working with moving machinery, and walking outdoors in darkness all become higher-risk activities.
• Genetic counsellingfor the m.1555A>G variant and first-degree relatives, particularly if the patient developed ototoxicity at unexpectedly low cumulative doses.

Differential diagnosis

Once vestibulotoxicity is established the picture is identical to bilateral vestibulopathy of any other cause. The clinical task is attribution rather than discrimination:

• Pre-existing bilateral vestibulopathy — the patient may have had subclinical loss before aminoglycoside exposure. Baseline DVA documentation is what distinguishes ototoxicity-caused loss from ototoxicity-revealed loss.
• Presbyvestibulopathy — older patients on aminoglycoside therapy may have age-related decline contributing to baseline thresholds.⁴¹
• CANVAS — the cerebellar-ataxia-neuropathy- vestibular-areflexia syndrome can mimic aminoglycoside vestibulotoxicity. Look for cerebellar oculomotor signs and chronic cough.²⁶
• Other ototoxic agents in combination — patients on multiple potentially ototoxic drugs (aminoglycoside + loop diuretic, aminoglycoside + cisplatin) carry additive risk.

Reading the report

A patient who developed bilateral vestibular loss during or after systemic aminoglycoside or platinum-based chemotherapy exposure, with vHIT gain < 0.6 bilaterally and/or caloric sum < 6 °/s per side, meets the Bárány criteria for bilateral vestibulopathy attributed to ototoxicity.^24,43 The DVA finding quantifies the functional consequence and provides the right outcome measure for both initial documentation and rehabilitation tracking. The single most useful clinical move at the time of diagnosis is the audit question: could this have been caught earlier with surveillance?⁴⁵